| Popis: |
The basic assumption of most of the classical and modern test theory models is that the item and test results are a product of only one latent dimension. Contemporary approaches in the field of test theory emphasize importance of analysis of potential sources of construct irrelevant variance (CIV) and bias which can invalidate general quality of test results (APA, 2014). Messick (1995) lists several broader sources of CIV: test characteristics, test administration conditions, respondent characteristics such as specific strategies and behaviours used in test situation. Test-taking strategies are important cognitive skills that strongly affect students’ performance in tests, but can be regarded as potential sources of CIV. In the context of test theory we still miss a theoretical frame for analysis and assessment of test taking strategies. Through several studies, we developed a battery of instruments aimed to measure some of the test taking strategies (overall speed of performance, persistence, tendency to guess at random, test wiseness) and behaviours (test anxiety, sensitivity to time constrain, general sensitivity to distractions). Our goal was to validate these instruments and assess correlations and the latent space defined with these test taking strategies, test behaviours and certain personality traits. Results were obtained on a sample of 600 university and 220 high school students. Preliminary results indicated satisfactory psychometric characteristics, including the latent structure, of the new instruments. Latent space analysis resulted with two wide orthogonal factors: adaptive characteristics (conscientiousness, low impulsivity, high positive perfectionism, slow work tempo or thoroughness, test wiseness, persistence) and non-adaptive characteristics (test anxiety, sensitivity to time constrain, general sensitivity to distractions, neuroticism, negative perfectionism). |
